One usage of these is in a secondary sort , which is used , when you
want the output values from Map sorted (within a key).

You implement a KeyComparator and tell mapreduce to use it to order
the keys using a composite key.

To ensure that during partioning & Grouping , all the records for a
key go to the same reducer
you need to define and set a Partitioner and ValueGroupingComparator.

search for Secondary Sort for more on this topic

regards

MJ

On Fri, Aug 17, 2012 at 12:00 PM, Anbarasan Murthy
<[email protected]> wrote:
> Hi,
>
> I have a question in mapreduce api.
>
> Would like to know the significance of the following items under jobconf
> class.
> ValueGroupingComparator
> KeyComparator
>
>
>
> Thanks,
> Anbu.



-- 
http://khangaonkar.blogspot.com/

Reply via email to